Design, synthesis, cholinesterase inhibition and molecular modelling study of novel tacrine hybrids with carbohydrate derivatives

Article abstract of DOI:10.1016/j.bmc.2018.10.003

A series of hybrids containing tacrine linked to carbohydrate-based moieties, such as D-xylose, D-ribose, and D-galactose derivatives, were synthesized by the nucleophilic substitution between 9-aminoalkylamino-1,2,3,4-tetrahydroacridines and the corresponding sugar-based tosylates. All compounds were found to be potent inhibitors of both acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) in the nanomolar IC₅₀ scale. Most of the D-xylose derivatives (6a-e) were selective for AChE and the compound 6e (IC₅₀ = 2.2 nM for AChE and 4.93 nM for BuChE) was the most active compound for both enzymes. The D-galactose derivative 8a was the most selective for AChE exhibiting an IC₅₀ ratio of 7.6 for AChE over BuChE. Only two compounds showed a preference for BuChE, namely 7a (D-ribose derivative) and 6b (D-xylose derivative). Molecular docking studies indicated that the inhibitors are capable of interacting with the entire binding cavity and the main contribution of the linker is to enable the most favorable positioning of the two moieties with CAS, PAS, and hydrophobic pocket to provide optimal interactions with the binding cavity. This finding is reinforced by the fact that there is no linear correlation between the linker size and the observed binding affinities. The majority of the new hybrids synthesized in this work do not violate the Lipinski's rule-of-five according to FAF-Drugs4, and do not demonstrated predicted hepatotoxicity according ProTox-II.

Full text of DOI:10.1016/j.bmc.2018.10.003

Accepted Manuscript
Design, synthesis, cholinesterase inhibition and molecular modelling study of
novel tacrine hybrids with carbohydrate derivatives
João Paulo Bizarro Lopes, Luana Silva, Gabriela da Costa Franarin, Marco
Antonio Ceschi, Diogo Seibert Lüdtke, Rafael Ferreira Dantas, Cristiane
Martins Cardoso de Salles, Floriano Paes Silva-Jr, Mario Roberto Senger,
Isabella Alvim Guedes, Laurent Emmanuel Dardenne
PII:
S0968-0896(18)31281-1
https://doi.org/10.1016/j.bmc.2018.10.003
BMC 14562
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
13 July 2018
5 October 2018
7 October 2018
Please cite this article as: Lopes, J.P.B., Silva, L., da Costa Franarin, G., Antonio Ceschi, M., Seibert Lüdtke, D.,
Ferreira Dantas, R., de Salles, C.M.C., Paes Silva-Jr, F., Roberto Senger, M., Alvim Guedes, I., Emmanuel Dardenne,
L., Design, synthesis, cholinesterase inhibition and molecular modelling study of novel tacrine hybrids with
carbohydrate derivatives, Bioorganic & Medicinal Chemistry (2018), doi: https://doi.org/10.1016/j.bmc.
2018.10.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Design, synthesis, cholinesterase inhibition and molecular modelling
study of novel tacrine hybrids with carbohydrate derivatives
João Paulo Bizarro Lopes a, Luana Silva a, Gabriela da Costa Franarin a, Marco Antonio Ceschi a*, Diogo
Seibert Lüdtke a, Rafael Ferreira Dantas b, Cristiane Martins Cardoso de Salles c, Floriano Paes Silva-Jr b,
Mario Roberto Senger b*, Isabella Alvim Guedes d, Laurent Emmanuel Dardenne d*
aInstituto de Química, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves, 9500, Campus do
b
Vale, 91501-970, Porto Alegre, RS, Brazil.
ldo Cruz, Av. Brasil, 4365, 21040-360, Rio de Janeiro,
c
RJ, Brazil. Instituto de Ciências Exatas, Universidade Federal Rural do Rio de Janeiro, BR 465, Km 7,
Campus Universitário, 23890-000, Seropédica, RJ, Brazil. dLaboratório Nacional De Computação
Científica-LNCC, Av. Getúlio Vargas, 333, Petrópolis, 25651-075, RJ, Brazil.
Abstract: A series of hybrids containing tacrine linked to carbohydrate-
based moieties, such as D-xylose, D-ribose, and D-galactose derivatives, were
synthesized by the nucleophilic substitution between 9-aminoalkylamino-
1,2,3,4-tetrahydroacridines and the corresponding sugar-based tosylates. All
compounds were found to be potent inhibitors of both acetylcholinesterase
(AChE) and butyrylcholinesterase (BuChE) in the nanomolar IC50 scale. Most of
the D-xylose derivatives (6a-e) were selective for AChE and the compound 6e
(IC50 = 2.2 nM for AChE and 4.93 nM for BuChE) was the most active
compound for both enzymes. The D-galactose derivative 8a was the most
selective for AChE exhibiting an IC50 ratio of 7.6 for AChE over BuChE. Only
two compounds showed a preference for BuChE, namely 7a (D-ribose
derivative) and 6b (D-xylose derivative). Molecular docking studies indicated
that the inhibitors are capable of interacting with the entire binding cavity and
the main contribution of the linker is to enable the most favorable positioning of
the two moieties with CAS, PAS, and hydrophobic pocket to provide optimal
interactions with the binding cavity. This finding is reinforced by the fact that
1

there is no linear correlation between the linker size and the observed binding
affinities. The majority of the new hybrids synthesized in this work do not violate
the Lipinski's rule-of-five according to FAF-Drugs4, and do not demonstrated
predicted hepatotoxicity according ProTox-II.
Keywords: Tacrine, carbohydrate, xylose, ribose, galactose, cholinesterases,
molecular modeling, Alzheimer.
*Corresponding authors;
E-mail: mceschi@iq.ufrgs.br, mario.senger@ioc.fiocruz.br, dardenne@lncc.br.
1. Introduction
Neurodegenerative disorders have complex pathological mechanisms
which result in progressive and irreversible loss of physiologically related
neuronal systems, consisting in one of the greatest challenges of medical
sciences. Known as the most common and devastating disorder of this type,
Alzheimer's Disease (AD) affects about 46.8 million people in the world, and it is
expected to increase by three fold by 2050, according to recent reports.[1,2]
Currently, there is no cure for AD and several hypotheses have been directing
research efforts, such as amyloid cascade, tau hyperphosphorylation, metal
imbalance, oxidative stress and cholinergic hypotheses, among others.[3] The
cholinergic hypothesis is the most related to the symptoms, which suggests that
the progressive cognitive impairment of AD results from a deficit in the
neurotransmitter acetylcholine (ACh), and consequent synaptic failure.[4,5] The
cholinesterase inhibitors (ChEI) promote the stimulation of the cholinergic
receptors, prolonging the availability of ACh in synaptic cleft, hence decreasing
the AD associated symptoms.[6]
2

The cholinergic theory remains the most effective therapeutic approach
for the symptomatic treatment of AD. Among the four drugs that have been
used so far, three are ChEI: donepezil (Aricept®), rivastigmine (Exelon®) and
galantamine (Razadyne®). Lately, memantine has been introduced into the
market as an alternative therapy acting as an antagonist of the NMDA receptor,
thereby protecting the nervous tissue against glutamate mediated excitotoxicity
with limit therapeutic effects.[7] Tacrine (chemically, 1,2,3,4-tetrahydroacridine,
THA, Figure 1) was, in 1993, the first drug approved for the treatment of AD,
and is one of the most efficient cholinesterase inhibitors, with high efficacy in
restoring ACh. However, it has been withdrawn from the market due to its
hepatotoxicity, mainly by metabolism to toxic quinone-type metabolites species
through the action of cytochrome P450 enzyme CYP1A2.[8]
Acetylcholinesterase (AChE, 3.1.1.7) belongs to the α/β hydrolase
protein fold and has been purified and crystallized in the monomeric, dimeric,
and tetrameric forms.[9] The active site is a narrow gorge with a length of
approximately 20 Å, which penetrates more than half-way into the enzyme,
containing two main subsites: the catalytic active site (CAS), located at the
bottom of the cavity, and the peripheral anionic site (PAS), near the entrance of
the gorge.[10,11] Besides the catalytic triad, composed by Ser200, His440 and
Glu3271, there are amino acids residues in the CAS strictly positioned to confer
high selectivity to acetyl esters, in order to preclude the hydrolysis of bulky
esters and to stabilize the acetate leaving group. These residues comprise,
respectively, the acyl pocket (Phe288, Phe290, Phe331) and the oxyanion hole
(Gly118, Gly119, Ala201).[12] One of the essential regions to the catalytic
process is the anionic subsite, composed mainly by Trp84, placed at the bottom
of the gorge, which plays a central role to attract ACh to CAS, interacting with
the ACh quaternary nitrogen.[10,12] Moreover, crystallographic and molecular
dynamics (MD) simulation data of TcAChE have pointed to the distance
alternation between Tyr442 and Trp84 as a backdoor, whose opening allows
1 All residues described in this text belong to AChE from Torpedo californica (TcAChE).
3

the substrate or product trafficking into or out of the cavity.[13] Several studies
have shown that AChE is not merely the enzyme which hydrolyzes ACh, but
also plays an essential role on the Aβ peptide assembly, acting as a chaperone
in Aβ folding.[14] It has been suggested that the PAS is the enzymatic region
responsible for this role, which has been strengthened by the fact that this
chaperone effect is sensitive only to drugs that block the PAS of the
AChE.[15,16]
Butyrylcholinesterase (BuChE, E.C. 3.1.1.8) is also involved in the
hydrolysis of the neurotransmitter ACh. It has been suggested that BuChE
activity rises in the elderly and in AD patients’ brains, while AChE activity
declines, indicating that BuChE plays a critical role for ACh hydrolysis in the late
stage of AD. Therefore, compounds that are able to inhibit both AChE and
BuChE might have better long-term therapeutic efficacy.[17-19] The structure of
human BuChE (HsBuChE) is very similar to HsAChE, sharing 65% amino acid
sequence homology and having the same kind of residues in the catalytic triad:
Ser198, His438, and Glu325 in HsBuChE.[19,20] The absence of bulky
aromatic amino acids in the PAS, especially in the acyl pocket of CAS, is the
main difference of BuChE when compared with AChE, being postulated as a
primary source of enzymatic selectivity.[12,21]
After tacrine has been approved for the treatment of AD, several tacrine-
based compounds have been studied boosted by their synthetic ease, highly
inhibitory profile and binding affinity with CAS and PAS of AChE. Despite its
hepatotoxicity, tacrine remains a reference and a good chemical scaffold to
design novel and more potent cholinergic prototypes with suppressed side
effects.[14,22] Recently, several 9-acridines derivatives were found as selective
cholinesterase inhibitors together with free radical scavenging, which makes
promising the use of the acridine scaffold to obtain multifunctional drugs for the
therapy of neurodegenerative diseases. [23] In 1996, Pang and co-workers
synthesized bis(n)-tacrine analogues linked by an alkylene chain which were up
to 10,000-fold more selective and 1,000-fold more potent than THA, inasmuch
as it has a double interaction in the enzyme, with both CAS and PAS.[24]
4

Recently, chiral bis(7)-tacrine analogues have been found to be potent and
selective AChE inhibitors in both enantiomeric series.[25] Moreover, the bis(7)-
tacrine (Figure 1) possesses remarkable neuroprotective activities, such as
inhibition of NO synthase, binding as antagonist of NMDA receptor and
inhibition of Aβ synthesis and fibrils formation.[26-28] Furthermore, in vivo
studies showed that bis(7)-tacrine was able to reach the AChE only 15 min after
the single-dose and intraperitoneal administration, indicating that it may easily
cross the blood–brain barrier.[29]
Figure 1. Structures of tacrine and bis(7)-tacrine.
Since then, bis(7)-tacrine has been the source of inspiration for the
design of new potential drugs for the treatment of AD and cognitive impairment.
The major strategy relies in the replacement of one tacrine moiety by a
pharmacophore or a biologically active nucleus that might be able to bind to
PAS.[14] In this context, several tacrine-based hybrids have been synthesized
in the latest years.[22,30-37] A series of tacrine-lophine and tacrine-tianeptine
hybrids, synthesized recently by our group, has demonstrated to be potent and
selective inhibitors of AChE and BuChE with IC50 in the nanomolar
concentration scale.[38,39]
Regarding the great efforts to delineate structural requirements for
optimal binding to the active site of AChE and BuChE, a good understanding of
their function and structure is essential to interweave salient details together
into a coherent predictive tool.[12] Since diverse experimentally determined
structures of AChE have been known and with the advancement of
computational chemistry, molecular modelling has been widely used to predict
5

the binding modes and affinity between the enzyme and inhibitors with different
scaffolds, guiding the optimization and development of new compounds. The
interactions of the THA moiety within the AChE binding site is highly consistent
in most of the compounds, being lodged in the CAS.[22] The THA binding mode
is mainly characterized by a π-π stacking with the Trp84 and Phe330 side
chain, and a hydrogen bond between the protonated aromatic nitrogen of the
quinoline ring and the carbonyl oxygen of the H440 main chain.[40-42] In
addition, it has been found that the secondary amino group present in many
potent inhibitors also makes important interactions such as hydrogen bonds
with different amino acid residues and/or water molecules.
Taking into consideration the potential of reported tacrine-based
molecules in the search of novel cholinesterase inhibitors, we designed new
hybrids connecting tacrine nucleus with sugar derivatives as showed in Figure
2. We believe that the biocompatibility of the carbohydrates might increase the
inhibitor's bioavailability as well as their permeability in biological barriers. In
addition, we also explore the intermolecular interactions between sugar moiety
and AChE, predicting the possible binding modes by molecular docking.
Besides the well-known hydrogen bonds between hydroxyl groups of the
carbohydrates and polar atoms in proteins, we also studied the less explored
CH/π interactions with aromatic residues, mainly tryptophan.[43-46] Moreover,
carbohydrates are multifunctionalizable structures, which allow the substitution
at various positions of the sugar ring and further structural modifications.
Recently, compounds containing carbohydrate scaffolds have been
studied as promising candidates for Alzheimer's therapy. O-flavonol glycosides
were synthesized by Mughal and coworkers and most of the derivatives were
potent inhibitors of AChE and BuChE with varying degree of IC50 values. [47]
Several purine nucleosides were synthesized by N-glycosylation and showed to
be potent inhibitors of both AChE and BuChE. [48,49] Iminossugar-based
compounds were synthesized and evaluated as cholinesterase inhibitors and
seven of the twenty-three compounds studied had IC50 values in the micromolar
range with significant selectivity for BuChE over AChE. In general, the
6

deprotection of the hydroxyl groups led to the reduction or absence of enzyme
inhibition.[50] Several studies with selenofuranosides have demonstrated
diverse activities associated with possible therapies for AD, such as antioxidant,
inflammation involvement, antidepressant-like effect and protection against
memory loss in mice.[51-55]
Figure 2. Carbohydrates scaffolds used in this work and the general structure of
the synthesized tacrine-carbohydrate hybrids.
2. Results and Discussion
2.1. Chemistry
The synthesis of tacrine-based hybrids required the previous preparation
of precursor 9-chloro-1,2,3,4-tetrahydroacridine (1) through the classical
Niementowski reaction between anthranilic acid and cyclohexanone mediated
by POCl3.[56] Treatment of the chloride with an excess of 1,n-diaminoalkanes
under reflux for 18 h provided 9-amino-1,2,3,4-tetrahydroacridines (2a-e) in
good yields (78%-95%), as depicted in Scheme 1. This step was carried out in
7

the presence of catalytic amount of KI.[57] The natural carbohydrates D-xylose,
D-ribose and D-galactose were straightforwardly converted into their tosylates,
respectively 3, 4 and 5, by a short synthetic sequence, according to reported
protocols (Scheme 2).[58-60]
Scheme 1. Preparation of tacrine derivative precursors (2a-e).
Scheme 2. Preparation of tosylates from natural sugars D-xylose, D-ribose and
D-galactose.
8

With the amines 2a-e and tosylates 3, 4 and 5 in hands, we turned our
attention to the connection of both nuclei tacrine and sugar derivatives through
the nucleophilic substitution of the tosylate leaving group. The general synthesis
of the novel tacrine-(carbohydrate-derived) hybrids is depicted in Scheme 3.
The reactants were mixed and solved in isopropanol in a sealed flask, and
tosylates 3, 4 and 5 were then converted into tacrine-based hybrids derived
from D-xylose, D-ribose and D-galactose, respectively. In this work, we opted for
the functionalization in the primary hydroxyl, keeping the anomeric position
protected. This choice was made to allow the stereochemical control, preclude
the oxidation in the physiological medium and enable further structural studies.
The acetonide group was chosen as the protecting group for the hydroxyl due to
their small size when compared to others, such as benzyl and acetyl.
9

Scheme 3. Synthesis of tacrine-(carbohydrate-derived) hybrids.
We also studied the deprotection of the hydroxyl groups using the most
common procedures to remove the acetonide group of carbohydrates, involving
the reaction with trifluoroacetic acid (TFA) in methanol or water, in short
reaction times. Unfortunately, the product was not obtained from anyone of the
evaluated conditions and we observed only a complex mixture, mainly by
degradation of the sugar moiety. We also intend to perform further studies to
enable the access of the deprotected compounds.
2.2. AChE and BuChE inhibition evaluation
The inhibitory activities against MmAChE and MmBuChE of new tacrine-
(carbohydrate-derived) hybrids, together with the reference compounds bis(7)-
tacrine and tacrine, are reported in Table 1. (For the corresponding inhibitor
titrating curves for IC50 determination, see supporting information, Figure S1).
All compounds were found to be potent inhibitors of both cholinesterases with
IC50 values in the nanomolar concentration scale. Moreover, six of nine
compounds were able to inhibit AChE in the same order of magnitude than the
bis(7)-tacrine. In general, the D-xylose derivatives (6a-e) exhibited AChE
modest selectivity and the compound 6e (n = 6, Entry 5) showed an IC50 of 2.2
nM against AChE and of 4.93 nM against BuChE, being the most potent
inhibitor of both AChE and BuChE in the assayed series of tacrine-carbohydrate
hybrids. Despite the higher potency, compound 6e showed the lower selectivity.
The compound 6b (n = 5, Entry 2) was the most selective for BuChE (IC50 ratio
of 0.14) and showed a good inhibitory activity for this enzyme (IC50 = 11.7 nM).
The D-galactose derived 8a (n = 6, Entry 5) was found to be the most selective
AChE inhibitor, exhibiting an IC50 ratio of 7.6 for AChE over BuChE.
Compounds derived from D-ribose (7a and 7b, Entries 6 and 7) showed slight
preference for BuChE.
10

Table 1. Inhibitory activity on AChE and BuChE, docking results against AChE (PDB ID 1ZGC)
and BuChE structures (5K5E), and the IC50 ratio of the studied compounds.
IC50 (nM) [confidence interval]
Docking Score*
IC50 ratio
BuChE/AChE
Entry
Compound
n
MmAChE
MmBuChE
TcAChE TcBuChE
6a
1
2
4
5
6
7
8
6
7
6
7
-
8.7 [7.3 – 10.5]
27.6 [19.3 - 39.7]
118.31
121.77
120.60
125.38
131.44
125.11
129.19
131.93
131.17
-
93.57
97.34
98.34
94.55
95.69
95.29
96.15
93.97
96.56
-
3.2
0.14
5.0
84.4 [63.6 –
6b
11.7 [8.7 - 15.9]
25.8 [18.5 - 36.0]
30.8 [20.4 - 46.6]
4.93 [4.0 - 6.1]
113.3]
6c
3
5.1 [3.6 – 7.6]
10.2 [7.2 – 14.5]
2.2 [1.8 – 2.7]
6d
4
3.0
6e
5
2.2
7a
7b
89.9 [62.4 – 131.0]
65.5 [45.2 – 93.7]
4.8 [3.2 – 7.4]
6
36.9 [21.8 - 62.5]
64.3 [50.2 - 82.5]
36.7 [22.9 - 58.7]
26.2 [19.1 - 36.1]
62.5 [44.0 - 89.0]
9.1 [6.6 - 12.5]
0.41
0.98
7.6
7
8a
8
8b
9
7.6 [5.6- 10.3]
3.4
34.3 [14.9 –
10
11
Tacrine
Bis(7)tacrine
1.8
107.8]
-
1.9 [1.5 – 2.4]
-
-
4.8
*Docking results obtained using the GOLD program with the ChemPLP scoring function. Higher score
value correspond to better protein-ligand binding strength.
2.3. Molecular Modelling
2.3.1. Redocking studies
We carried out redocking experiments for the AChE and BuChE
structures used in this work to validate the docking protocol adopted (i.e., GOLD
docking program with the ChemPLP scoring function). In all four complexes
(three structures of AChE and one structure of BuChE), our docking
methodology was able to predict the top-ranked pose close to the experimental
conformation observed for the reference ligand present on each crystallographic
structure (Figure 3), even for the highly flexible AChE inhibitors. In general,
11

docking studies are considered successful when the root-mean-square
deviation (RMSD) between the docked pose and a reference conformation is
less than 2 Å.
Figure 3. Top-ranked poses predicted by GOLD::ChemPLP for each AChE
(PDB IDs 2CKM, 1Q84 and 1ZGC) and BuChE (5K5E) structure. The predicted
and experimentally observed binding mode are highlighted in yellow and green,
respectively.
2.3.2. Ensemble docking
According to the ensemble docking strategy adopted herein for the
AChE, the top-scored binding mode of all compounds predicted by the
ChemPLP scoring function was found for the 1ZGC conformation (Table S1). In
our docking strategy, we were able to find three highly potent AChE inhibitors
as the top-ranked compounds. For BuChE, all compounds obtained significantly
worst scores, and no correlation between them and the experimental binding
data was found. For instance, 6b exhibit selectivity against BuChE over AChE
(IC50 BuChE/AChE ratio of 0.14), but ChemPLP predicted better score for AChE
than BuChE. This might indicate that further experiments are needed to explore
12

additional BuChE conformations, since there are no BuChE crystal structures
available with such large compounds.
2.3.3. Contribution of the carbohydrate moiety (compounds with linker size n =
6)
We analyzed the influence of the carbohydrate moiety in the predicted
binding mode and affinity against AChE for the 6-carbon linker compounds 8a,
7a and 6c (Table 1). These three compounds exhibit similar overall binding
modes (Figure 4A): (i) the tacrine moiety interact at the bottom of the gorge
through the conserved interactions already described in the literature; (ii) the
linker is located at the middle of the gorge performing hydrophobic contacts
mainly with Tyr331; (iii) the carbohydrate moiety interacts in the PAS region.
Although ChemPLP is not able to compute specific contributions from CH/π
interactions and correctly rank these derivatives, the binding modes suggested
by docking indicated this kind of interactions between the carbohydrates and
the Trp279 from PAS (Figures 4B and 5). Furthermore, it is possible that the
positive charge of the protonated nitrogen atom of the linker is imparted on the
neighboring carbon atoms, increasing the strength of the CH/π interactions and
helping to stabilize the sugar moiety with PAS. Many theoretical and
experimental studies highlighted the importance of CH/π interactions in
molecular recognition and binding affinity, with tryptophan being the most
frequent aromatic residue found in contact with carbohydrates.[43,46,61,62].
Despite the potent compound 6c obtained a relatively poor score (120.6), it was
found as one of the most potent AChE inhibitors of the series (IC50 = 5.1 nM). In
addition to the interaction with Trp279, the binding mode predicted for this
compound exhibits an intramolecular interaction between the protonated
nitrogen atom and the hydroxyl group from xylose, at the same time as this
oxygen atom is within hydrogen bond distance from the Tyr70 hydroxyl group
(2.6 Å). For the compound 8a, besides the CH/π interaction with Trp279, the
13

acetonide group also interact with a hydrophobic pocket formed by Ile287 and
Phe331 (Figure 4B). These nonpolar contacts are also observed for the less
potent inhibitor 7a. The reference ligand bound in the 1ZGC complex has a
carbonyl oxygen atom in the same region interacting with the Phe285 backbone
NH atom through a hydrogen bond (3.2 Å) (Figure 4C).[63] However, in the
compounds 8a and 7a, the nearest oxygen atoms of the acetonide group to the
Phe288 NH (4.0 and 5.0 Å, respectively) are more distant than the limiting
criteria for hydrogen bond formation. Thus, further structural modifications
guided by molecular docking could even improve the potency of 8a and 7a
against AChE.
(Insert
Figure
4
here)
14

Figure 4. Docking results for the compounds 6c (pink), 7a (yellow) and 8a
(green) against AChE (PDB ID 1ZGC). A) Predicted binding mode of the
compounds interacting with the entire AChE gorge. B) Interaction between the
compounds and a methyl group from acetonide with the hydrophobic pocket
formed by Ile287 and Phe331. C) Superposition of the most potent inhibitors 8a
and 7a with the co-crystallized ligand in the complex 1ZGC. Hydrogen bonds
are represented as yellow dashes.
Figure 5. Predicted interactions of the xylose, ribose and galactose moieties of
the compounds 6c, 7a and 8a (green) with Trp279 from PAS from AChE (PDB
ID 1ZGC). Possible CH/π interactions are represented as pink dashed lines and
the carbohydrate moieties are represented as sticks.
2.3.4. Influence of the linker size – xylose derivatives
According to the ChemPLP predictions for AChE, in accordance with the
binding data, the linker size of eight carbon atoms led to the most potent xylose
derivative 6e (Table 1, IC50 = 2.2 nM, ChemPLP score = 131.4). Strikingly, the
compound 6b differs only in one carbon in the linker when compared with 6a
and 6c, but exhibits significantly distinct IC50 value against AChE. Unfortunately,
we could not find correlations between the observed binding affinities and the
15

predicted binding scores for the xylose-derivatives that could explain the
influence of the linker size to the potency of these compounds.
According to the docking results, the most potent compound 6e is able to
interact with both pockets (Figure 6A): the xylose moiety interacts with the
hydrophobic binding site, mainly with Ile287, while the protonated nitrogen atom
and neighboring carbon atom interact with PAS, possibly through cation-pi
interactions. Additionally, the hydroxyl group from the xylose moiety is exposed
to the solvent, thus avoiding any polar desolvation cost. The eight-carbon linker
seems to be an adequate long linker, allowing a more efficient blockage of the
AChE gorge due to the interaction of the sugar moiety with both pockets in the
entrance of the cavity (Figure 6B). The less potent xylose-derivative, compound
6b, mainly interacts with the entrance of the gorge through two hydrogen bonds
with Tyr70 and Tyr121 (Figure 6C and 6D). The compounds 6a and 6c, which
differs from the 6b structure with a single carbon atom in the linker, were
predicted to assume significant distinct positions for the xylose moiety. The
shortest linker of 6a allowed for a hydrogen bond between the protonated
nitrogen atom and the Tyr121 side chain (2.8 Å), while the xylose group is
located near the hydrophobic pocket (Figure 6C). On the other hand, the
compound 6c was able to interact with Trp279 and is hydrogen bonded with
Tyr70 side chain (2.6 Å) (Figure 6D), probably stabilized by the intramolecular
hydrogen bond discussed elsewhere (Figure 4B).
16

Figure 6. Docking results of the compounds 6a (purple), 6b (light blue), 6c
(pink), 6d (dark green) and 6e (grey) against AChE (PDB ID 1ZGC). Hydrogen
bonds are represented as yellow dashes.
In this work, the synthesis was directed to compounds containing the
hexa- and heptamethylenic linker, which allow a optimal distance between CAS
and PAS of the AChE, as well as bis(7)-tacrine.[24] The D-xylose derivatives
were chosen to have their scope expanded from n = 4 to n = 8 to evaluate the
influence of the linker size. According to our docking results as discussed
above, shorter linkers (n = 4 and n = 5) might not furnish the most favorable
17

positioning of sugar moiety in the PAS, with difficulty to establish important
interactions. On the other hand, in spite of the larger linker (n = 8) provided a
more efficient blockage of the AChE cavity entrance by the sugar moiety, the
enzymatic selectivity was lost, possibly by the greater flexibility and ability of the
molecule to adjust in both AChE and BuChE active sites.
2.3.5. ADME-Tox prediction
We submitted the tacrine-carbohydrate hybrids synthesized in this study
through the free FAF-Drugs4 webserver to predict their drug-like properties.
[64,65] The Lipinski’s “Rule-of-five” concerns a set of molecular descriptors
widely used to predict potential oral drugs, such as: molecular weight (MW) ≤
500 Da, lipophilicity assessed by the ClogP ≤ 5 (the logarithm of the partition
coefficient between water and octanol), number of hydrogen bond donors
(HBD) and acceptors (HBA), respectively ≤ 5 and ≤ 10.[66] According to the
predictions, five derivatives do not violate the Lipinski's rules (6a-d and 7a), and
four compounds only violate the MW rule (6e, 7a, and 7b). Therefore, all
carbohydrate derivatives were predicted to have good oral absorption.
Table 2. Drug likeness of the carbohydrate derivatives, according to the Lipinski’s rules evaluated by the
FAFDrugs4 webserver.
Entry
Compound
MW
HBD
HBA
7
ClogP
2.96
3.32
3.68
4.22
4.76
3.66
4.20
4.07
4.61
Violations
6a
6b
6c
6d
6e
7a
7b
8a
8b
1
2
3
4
5
6
7
8
9
447.61
461.64
475.66
489.69
503.72
489.69
503.72
545.75
559.78
4
4
4
4
4
3
3
3
3
0
0
0
0
1
0
1
1
1
7
7
7
7
7
7
8
8
18

2.3.6. Hepatotoxicity prediction
To verify a possible hepatotoxic effect of the compounds, we used
ProTox-II, a webserver for the prediction of small molecules toxicity [67]. We
utilized the hepatotoxicity predictor and the main results are expressed in Table
S2 . All of the nine compound synthetized in this work were predicted to have a
5 fold higher LD50 than tacrine. Furthermore, tacrine bellows to a higher
predicted toxicity class than the tacrine-carbohydrate compounds. Also, tacrine
was predicted to mediate its toxicity by interacting with Amine Oxidase A, fact
that was not observed for the new compound class described in this work.
3. Conclusion
In this work, we synthesized a series of novel tacrine hybrids with
natural-based D-xylose, D-ribose and D-galactose with high AChE and BuChE
inhibitory activity in the nanomolar concentration scale. The synthesis of the
hybrids is accomplished by the displacement reaction between electrophilic
carbohydrate derivatives in the tosylate form with the appropriated 9-
alkylamino-1,2,3,4-tetrahydroacridines.
All compounds were found to be potent inhibitors of both cholinesterases
with IC50 values in the nanomolar concentration scale. Moreover, six of nine
compounds were able to inhibit AChE in the same order of magnitude than the
bis(7)-tacrine. In general, the D-xylose and D-galactose derivatives exhibited
AChE modest selectivity whereas the hybrids derived from D-ribose showed a
slight preference for BuChE. Our ensemble docking results indicate that the
main contribution of the linker is to enable the correct positioning of the two
moieties with CAS, PAS, and hydrophobic pocket to provide optimal interactions
with the binding cavity, providing an efficient blockage of the binding site. This
finding is reinforced by the fact that there is no linear correlation between the
19

linker size and the observed binding affinities. Also, the sugar moieties are
stabilized in the PAS region through cation-π and CH/π interactions with
Trp279, hydrophobic interactions from their acetonide protecting groups and
hydrogen bonds.
Overall, the compounds synthesized in this work, with tacrine connected
to a carbohydrate nucleus, represent a novel approach in the literature. The
desired hybrids were obtained from simple starting materials in just a few steps
and showed high inhibitory potency. Most of them were predicted to obey the
Lipinski's rules and do not demonstrated predicted hepatotoxicity, representing
new promising prototypes for the treatment of AD.
4. Experimental protocols
4.1. Chemistry
1H NMR and 13C NMR spectra were recorded in CDCl3 solution on a
Varian VNMRS 300 MHz spectrometer. The assignment of chemical shifts is
13
1
1
1
13
based on standard NMR experiments (1H; C; H, H-COSY; H, C-HMQC).
Chemical shifts (δ) are given in part per million from the peak of
1
tetramethylsilane (δ = 0.00 ppm) as internal standard in H NMR or from the
solvent peak of CDCl3 (δ = 77.23 ppm) in 13C NMR. NMR multiplicities are given
as s (singlet), d (doublet), dd (doublet of doublets), ddd (doublet of doublet of
doublets), t (triplet), m (multiplet) or br (broad) and coupling constants (J) are
given in Hz. IR spectra were recorded on a Varian 640-IR spectrometer in KBr
pellets. Specific rotations were measure on a Jasco P-2000 polarimeter with a
1.0 mL cell at a temperature of 20 ºC. High Resolution Mass Spectrometry with
Eletrospray Ionization (HRMS-ESI) data on the positive mode was collected on
a UHPLC-QTOF/MS Bruker Impact II. Samples were infused from a 100 mL
Hamilton syringe at flow rate range from 5 to 10 mL/min, depending on the
sample. The instrument settings were the following: capillary voltage 3000 V,
20

cone voltage 33 V, extraction cone voltage 2.5 V, desolvation gas temperature
100 ºC. High performance liquid chromatography (HPLC) was performed in a
Dionex Ultimate 3000 apparatus, equipped with column Shim-pack CLC-ODS
(M) (250 mm x 4,6 mm, 5 µm). Stock solutions were prepared by dissolution of
1 mg from each compound in 1 mL of acetonitrile (HPLC grade Vetec), from
which the working solutions were prepared by diluting 100 μL of each sample in
1 mL of acetonitrile. The working solutions were filtered in PVDF Flowsupply
filter with 0,22 µm of porosity being followed by HPLC-UV analyses at 25 ºC,
injection's volume of 10 µL, isocratic elution with acetonitrile: water (60: 40), and
rate of flow of 1 mL.min-1, 10 min of acquisition time and 270 nm. The absence
of significant peaks besides the compound peak evidence of sample high purity
of compounds. Nitrogen was used as the desolvation gas. Methanol (Tedia
HPLC grade) was used as solvent for the analysed samples and filtered prior to
injection. Purification by column chromatography was carried out on silica gel
60 Å (70-230 mesh) and analytical thin layer chromatography (TLC) was
conducted on aluminium plates with 0.2 mm of silica gel 60F-254 (Macherey-
Nagel). Solvents were obtained from Tedia and Nuclear, and reagents were
purchased from Sigma-Aldrich, Acros Organics and TCI. The compounds 1
[56], 2a-e [57] and tosylates 3, 4 and 5 [58-60], were prepared according to the
previous literature procedure.
4.1.1. General procedure for the preparation of tacrine-(carbohydrate-derivative)
hybrids
A mixture containing 0.3 mmol of tosylate from D-xylose (3), D-ribose (4)
or D-galactose (5) and 0.6 mmol of 9-alkylamino-1,2,3,4-tetrahydroacridines
(2a-e) was dissolved in isopropyl alcohol (1.0 mL) and was maintained under
stirring at 83 ºC during 72 hours in a sealed flask. After this time, the reaction
was quenched by saturated NaHCO3 (20 mL), washed with brine (20 mL) and
water (20 mL) and the product was extracted with dichloromethane. The organic
21

layer was dried by Na2SO4, the solvent was removed under vacuum and the
crude product was purified by column chromatography, eluting with hexane:
ethyl acetate 1: 1 until remove remaining tosylate, follow by chloroform:
methanol 98: 2 to give the desired product. Samples used in biological assays
showed high purity as observed by NMR and HPLC analyzes.
4.1.1.1.
N1-(5-Deoxy-1,2-O-isopropylidene-α-D-xylofuranose)-5-(N4-(1,2,3,4-
tetrahydroacridin-9-yl)butane-1,4-diamine) (6a).
Intermediate 2a and tosylate 3 were reacted according to general procedure to
20
give the desired product 6a as a yellow oil (54% yield); [α]D = +63.6º (c 0.154,
CH2Cl2); IR (KBr) νmax/cm-1: 3397, 2920, 2853, 1562, 1499, 1372, 1072, 1009,
1
752; H-NMR (CDCl3, 300 MHz) δ 7.97 – 7.87 (m, 2H), 7.55 (ddd, J = 8.4, 6.9,
1.2 Hz, 1H), 7.39 – 7.31 (m, 1H), 5.95 (d, J = 3.7 Hz, 1H), 4.49 (d, J = 3.6 Hz,
1H), 4.28 (d, J = 2.8 Hz, 1H), 4.21 (d, J = 2.2 Hz, 1H), 3.98 (br, 2H), 3.56 – 3.45
(m, 2H), 3.38 (dd, J = 12.9, 3.5 Hz, 1H), 3.11 – 3.02 (m, 2H), 2.95 (dd, J = 13.0,
1.3 Hz, 1H), 2.75 – 2.65 (m, 2H), 2.65 – 2.51 (m, 1H), 1.98 – 1.86 (m, 4H), 1.77
– 1.53 (m, 4H), 1.50 – 1.46 (s, 3H), 1.34 – 1.24 (m, 4H); 13C-NMR (CDCl3, 75
MHz) δ 158.5 (C0), 150.9 (C0), 147.3 (C0), 128.6 (C0), 123.9 (CH), 122.9 (CH),
122.6 (CH), 120.3 (CH), 116.2 (C0), 111.6 (C0), 105.2 (CH), 86.2 (CH), 78.4
(CH), 77.1 (CH), 49.3 (2CH2), 48.7 (CH2), 34.0 (CH2), 29.5 (CH2), 27.2 (CH3),
27.0 (CH3), 26.3 (CH2), 24.9 (CH2), 23.1 (CH2), 22.9 (CH2); HRMS-ESI: calcd
for [M-H]+ 442.2700, found 442.2695.
4.1.1.2.
N1-(5-Deoxy-1,2-O-isopropylidene-α-D-xylofuranose)-5-(N5-(1,2,3,4-
tetrahydroacridin-9-yl)pentane-1,5-diamine) (6b).
Intermediate 2b and tosylate 3 were reacted according to general procedure to
20
give the desired product 6b as a yellow oil (70% yield); [α]D = +60.9º (c 0.328,
1
CH2Cl2); IR (KBr) νmax/cm-1: 3381, 2940, 2843, 1572, 1487, 1072, 998, 757; H-
NMR (CDCl3, 300 MHz) δ 7.94 (t, J = 7.8 Hz, 2H), 7.61 – 7.52 (m, 1H), 7.40 –
7.30 (m, 1H), 5.95 (d, J = 3.7 Hz, 1H), 4.49 (d, J = 3.6 Hz, 1H), 4.28 (d, J = 2.8
22

Hz, 1H), 4.20 (d, J = 2.3 Hz, 1H), 4.04 (br, 2H), 3.49 (t, J = 7.2 Hz, 2H), 3.38
(dd, J = 13.0, 3.5 Hz, 1H), 3.07 (s, 2H), 3.00 – 2.89 (m, 1H), 2.76 – 2.65 (m,
2H), 2.61 – 2.48 (m, 1H), 1.97 – 1.88 (m, 4H), 1.74 – 1.61 (m, 2H), 1.58 – 1.35
13
(m, 7H), 1.35 – 1.21 (m, 4H); C-NMR (CDCl3, 75 MHz) δ 157.9 (C0), 150.9
(C0), 146.7 (C0), 128.4 (CH), 128.0 (CH), 123.6 (CH), 122.9 (CH), 119.9 (C0),
115.6 (C0), 111.3 (C0), 105.0 (CH), 85.9 (CH), 77.9 (CH), 77.0 (CH), 49.2 (CH2),
49.1 (CH2), 48.4 (CH2), 33.6 (CH2), 31.4 (CH2), 29.2 (CH2), 26.8 (CH3), 26.1
(CH3), 24.7 (CH2), 24.4 (CH2), 22.7 (CH2), 22.6 (CH2); HRMS-ESI: calcd for [M-
H]+ 456.2857; found 456.2855.
4.1.1.3.
N1-(5-Deoxy-1,2-O-isopropylidene-α-D-xylofuranose)-5-(N6-(1,2,3,4-
tetrahydroacridin-9-yl)hexane-1,6-diamine) (6c).
Intermediate 2c and tosylate 3 were reacted according to general procedure to
give the desired product 6c as a yellow oil (86% yield); [α]D20 = -65.0º (c 0.303,
CH2Cl2); IR (KBr) νmax/cm-1: 3369, 3928, 2847, 2345, 1502, 1081, 1001, 760,
1
659; H-NMR (CDCl3, 300 MHz) δ 7.98 – 7.86 (m, 2H), 7.54 (ddd, J = 8.3,
6.9,1.3 Hz, 1H), 7.35 (ddd, J = 8.3, 6.8, 1.3 Hz, 1H), 5.95 (d, J = 3.6 Hz, 1H),
4.49 (d, J = 3.5 Hz, 1H), 4.28 (d, J = 2.6 Hz, 1H), 4.19 (m, 1H), 3.95 (br, 1H),
3.54 – 3.42 (m, 2H), 3.36 (dd, J = 13.0, 3.3 Hz, 1H), 3.06 (s, 2H), 2.93 (d, J =
11.7 Hz, 1H), 2.68 (s, 2H), 2.64 – 2.40 (m, 2H), 1.97 – 1.84 (m, 4H), 1.72 – 1.58
13
(m, 2H), 1.53 – 1.22 (m, 12H); C-NMR (CDCl3, 75 MHz) δ 158.4 (C0), 150.8
(C0), 147.4 (C0), 128.6 (CH), 128.3 (CH), 123.6 (CH), 122.9 (CH), 120.2 (C0),
115.8 (C0), 111.4 (C0), 105.1 (CH), 85.8 (CH), 78.1 (CH), 77.0 (CH), 49.4 (CH2),
48.5 (CH2), 34.0 (CH2), 31.7 (CH2), 29.5 (CH2), 26.9 (1CH2 and 1CH3), 26.7
(CH3), 26.2 (CH2), 24.8 (CH2), 23.1 (CH2), 22.8 (CH2); HRMS-ESI: calcd for [M-
H]+ 470.3013; found 470.3014.
4.1.1.4.
tetrahydroacridin-9-yl)heptane-1,7-diamine) (6d).
N1-(5-Deoxy-1,2-O-isopropylidene-α-D-xylofuranose)-5-(N7-(1,2,3,4-
23

Intermediate 2d and tosylate 3 were reacted according to general procedure to
give the desired product 6d as a yellow oil (81% yield); [α]D20 = +87.7º (c 0.106,
1
CH2Cl2); IR (KBr) νmax/cm-1 3423, 2920, 2840, 1621, 1066, 604; H-NMR
(CDCl3, 300 MHz) δ 8.02 – 7.85 (m, 2H), 7.61 – 7.48 (m, 1H), 7.41 – 7.30 (m,
1H), 5.95 (d, J = 3.6 Hz, 1H), 4.49 (d, J = 3.6 Hz, 1H), 4.28 (d, J = 2.7 Hz, 1H),
4.20 (s, 1H), 4.00 (br, 1H), 3.56 – 3.44 (m, 2H), 3.39 (dd, J = 12.9, 3.3 Hz, 1H),
3.07 (s, 2H), 2.94 (d, J = 13.1 Hz, 1H), 2.79 – 2.45 (m, 4H), 1.92 (s, 4H), 1.65
13
(s, 2H), 1.57 – 1.23 (m, 14H). ; C-NMR (CDCl3, 75 MHz) δ 158.1 (C0), 150.8
(C0), 147.2 (C0), 128.3 (2CH), 123.5 (CH), 122.9 (CH), 120.1 (C0), 115.6 (C0),
111.3 (C0), 105.0 (CH), 86.0 (CH), 77.9 (CH), 77.0 (CH), 49.4 (CH2), 49.3
(CH2), 48.5 (CH2), 33.8 (CH2), 31.6 (CH2), 29.4 (CH2), 29.0 (CH2), 26.9 (CH3),
26.7 (CH3), 26.1 (CH2), 24.7 (CH2), 23.0 (CH2), 22.7 (CH2); HRMS-ESI: calcd
for [M-H]+ 484.3170; found 484.3164.
4.1.1.5.
N1-(5-Deoxy-1,2-O-isopropylidene-α-D-xylofuranose)-5-(N8-(1,2,3,4-
tetrahydroacridin-9-yl)octane-1,8-diamine) (6e).
Intermediate 2e and tosylate 3 were reacted according to general procedure to
give the desired product 6e as a yellow oil (29% yield); [α]D20 = -12.5º (c 0.990,
CH2Cl2); IR (KBr) νmax/cm-1: 3324, 2914, 2846, 1573, 1561, 1492, 1370, 1068,
1
1000, 754, 726; H-NMR (CDCl3, 300 MHz) δ 8.01 – 7.89 (m, 2H), 7.59 – 7.51
(m, 1H), 7.40 – 7.30 (m, 1H), 5.95 (d, J = 3.7 Hz, 1H), 4.48 (d, J = 3.7 Hz, 1H),
4.28 (d, J = 2.9 Hz, 1H), 4.20 (d, J = 2.9 Hz, 1H), 4.04 (br, 1H), 3.53 – 3.43 (m,
2H), 3.39 (dd, J = 13.0, 3.5 Hz, 1H), 3.11 – 3.00 (m, 2H), 2.94 (dd, J = 13.0, 1.3
Hz, 1H), 2.78 – 2.45 (m, 4H), 1.96 – 1.82 (m, 4H), 1.70 – 1.55 (m, 2H), 1.52 –
13
1.10 (m, 16H); C-NMR (CDCl3, 75 MHz) δ 158.3 (C0), 151.2 (C0), 147.2 (C0),
128.7 (CH), 128.6 (CH), 123.9 (CH), 123.1 (CH), 120.2 (C0), 115.7 (C0), 111.6
(C0), 105.3 (CH), 86.2 (CH), 78.4 (CH), 77.1 (CH), 49.7 (2CH2), 49.6 (CH2),
48.7 (CH2), 33.9 (CH2), 31.9 (CH2), 29.9 (CH2), 29.7 (CH2), 29.5 (CH2), 29.4
(CH2), 27.2 (CH2), 27.0 (2CH3), 26.3 (CH2), 24.9 (CH2), 23.2 (CH2), 22.9 (CH2);
HRMS-ESI: calcd for [M-H]+ 498.3326, found 498.3329.
24

4.1.1.6. N1-(1-Methyl-5-deoxy-2,3-O-isopropylidene-β-D-ribofuranoside)-5-(N6-
(1,2,3,4-tetrahydroacridin-9-yl)hexane-1,6-diamine) (7a).
Intermediate 2c and tosylate 4 were reacted according to general procedure to
20
give the desired product 7a as a yellow oil (35% yield); [α]D = -54.7º (c 1.10,
1
CH2Cl2); IR (KBr) νmax/cm-1: 3386, 2926, 2851, 1566, 1089, 867, 760; H-NMR
(CDCl3, 300 MHz) δ 7.98 – 7.87 (m, 2H), 7.57 – 7.51 (m, 1H), 7.37 – 7.30 (m, 1H),
4.96 (s, 1H), 4.62 (d, J = 6.0 Hz, 1H), 4.57 (d, J = 6.0 Hz, 1H), 4.30 (t, J = 6.7 Hz, 1H),
3.93 (s, 1H), 3.47 (t, J = 7.1 Hz, 2H), 3.33 (s, 3H), 3.09 - 2.99 (m, 3H), 2.78 – 2.66 (m,
4H), 2.66 – 2.51 (m, 2H), 1.98 – 1.83 (m, 5H), 1.73 – 1.59 (m, 3H), 1.55 – 1.28 (m,
10H); 13C-NMR (CDCl3, 75 MHz) δ 158.4 (C0), 151.1 (C0), 147.4 (C0), 128.7
(CH), 128.6 (CH), 123.8 (CH), 123.0 (CH), 120.3 (C0), 115.9 (C0), 112.5 (C0),
109.9 (CH), 86.4 (CH), 85.6 (CH), 82.9 (CH), 55.3 (CH2), 53.3 (CH2), 49.9
(CH2), 49.6 (CH2), 34.1 (CH2), 31.9 (CH2), 30.2 (CH2), 27.2 (CH2), 27.1 (CH2),
26.7 (CH3), 25.2 (CH3), 24.9 (CH2), 23.2 (CH2), 22.9 (CH2); HRMS-ESI: calcd
for [M-H]+ 484.3170, found 484.3170.
4.1.1.7. N1-(1-Methyl-5-deoxy-2,3-O-isopropylidene-β-D-ribofuranoside)-5-(N7-
(1,2,3,4-tetrahydroacridin-9-yl)heptane-1,7-diamine) (7b).
Intermediate 2d and tosylate 4 were reacted according to general procedure to
give the desired product 7b as a yellow oil (29% yield); [α]D20 = -37.9º (c 0.287,
CH2Cl2); IR (KBr) νmax/cm-1: 3370, 2921, 2846, 1554, 1493, 1371, 1098, 863,
1
750; H-NMR (CDCl3, 300 MHz) δ 7.98 – 7.86 (m, 2H), 7.59 – 7.49 (m, 1H),
7.33 (t, J = 7.2 Hz, 1H), 4.95 (s, 1H), 4.62 (d, J = 6.0 Hz, 1H), 4.57 (d, J = 6.0
Hz, 1H), 4.30 (t, J = 6.9 Hz, 1H), 3.93 (br, 2H), 3.47 (t, J = 7.2 Hz, 2H), 3.31 (s,
3H), 3.10 – 2.97 (m, 2H), 2.73 – 2.64 (m, 4H), 2.64 – 2.52 (m, 2H), 1.98 – 1.84
(m, 4H), 1.76 – 1.58 (m, 4H), 1.56 – 1.19 (m, 12H); 13C-NMR (CDCl3, 75 MHz)
δ 158.6 (C0), 151.0 (C0), 147.6 (C0), 128.9 (CH), 128.4 (CH), 123.8 (CH), 123.0
(CH), 120.4 (C0), 116.0 (C0), 112.5 (C0), 109.9 (CH), 86.4 (CH), 85.6 (CH), 82.9
(CH), 55.3 (CH2), 53.3 (CH2), 50.0 (CH2), 49.7 (CH2), 34.2 (CH2), 31.9 (CH2),
25

30.2 (CH2), 29.5 (CH2), 27.4 (CH2), 27.1 (CH2), 26.7 (CH3), 25.2 (CH3), 25.0
(CH2), 23.3 (CH2), 23.0 (CH2); HRMS-ESI: calcd for [M-H]+ 498.3326, found
498.3327.
4.1.1.8. N1-(6-Deoxy-1,2:3,4-di-O-isopropylidene-α-D-galactopyranose)-5-(N6-
(1,2,3,4-tetrahydroacridin-9-yl)hexane-1,6-diamine) (8a).
Intermediate 2c and tosylate 5 were reacted according to general procedure to
20
give the desired product 8a as a yellow oil (58% yield); [α]D = -26.0º (c 1.09,
CH2Cl2); IR (KBr) νmax/cm-1: 3338, 2982, 2928, 2852, 1561, 1383, 1205, 1068,
1
993, 768, 734, 522; H-NMR (CDCl3, 300 MHz) δ 7.98 – 7.85 (m, 2H), 7.57 –
7.48 (m, 1H), 7.36 – 7.27 (m, 1H), 5.54 (d, J = 5.1 Hz, 1H), 4.59 (dd, J = 7.9, 2.3
Hz, 1H), 4.31 (dd, J = 5.1, 2.3 Hz, 1H), 4.17 (dd, J = 7.9, 1.8 Hz, 1H), 3.98 (br,
1H), 3.96 – 3.85 (m, 1H), 3.52 – 3.37 (m, 2H), 3.05 (s, 2H), 2.95 – 2.44 (m, 6H),
1.97 – 1.81 (m, 4H), 1.71 – 1.56 (m, 2H), 1.55 – 1.21 (m, 18H); 13C-NMR
(CDCl3, 75 MHz) δ 158.2 (C0), 150.8 (C0), 147.3 (C0), 128.5 (CH), 128.3 (CH),
123.5 (CH), 122.9 (CH), 120.1 (C0), 115.7 (C0), 109.1 (C0), 108.4 (C0), 96.4
(CH), 72.0 (CH), 70.8 (CH), 70.5 (CH), 66.7 (CH), 49.6 (CH2), 49.4 (CH2), 49.2
(CH2), 33.9 (CH2), 31.7 (CH2), 29.9 (CH2), 27.1 (CH2), 26.8, 26.1, 26.0, 24.9,
24.8, 24.3, 23.0, 22.7 (4CH2 and 4CH3); HRMS-ESI: calcd for [M-H]+ 540.3432;
found 540.3435.
4.1.1.9. N1-(6-Deoxy-1,2:3,4-di-O-isopropylidene-α-D-galactopyranose)-5-(N7-
(1,2,3,4-tetrahydroacridin-9-yl)heptane-1,7-diamine) (8b).
Intermediate 2d and tosylate 5 were reacted according to general procedure to
give the desired product 8b as a yellow oil (34% yield); [α]D20 = -19.8º (c 0.136,
CH2Cl2); IR (KBr) νmax/cm-1: 3355, 2928, 2848, 1555, 1499, 1361, 1201, 1064,
991, 758; 1H-NMR (CDCl3, 300 MHz) δ 7.96 (t, J = 6.0 Hz, 2H), 7.57 (t, J = 7.7
Hz, 1H), 7.40 – 7.30 (m, 1H), 5.54 (d, J = 5.0 Hz, 1H), 4.60 (d, J = 7.7 Hz, 1H),
4.31 (d, J = 5.0 Hz, 1H), 4.19 (d, J = 8.0 Hz, 1H), 3.92 (d, J = 4.6 Hz, 1H), 3.51
(dd, J = 12.9, 6.3 Hz, 2H), 3.09 (s, 2H), 2.89 (dd, J = 12.9, 8.7 Hz, 1H), 2.81 –
26

2.51 (m, 5H), 2.08 – 1.83 (m, 5H), 1.75 – 1.57 (m, 2H), 1.59 – 1.22 (m, 20H);
13C-NMR (CDCl3, 75 MHz) δ 158.0 (C0), 151.4 (C0), 146.9 (C0), 128.8 (CH),
128.2 (CH), 123.9 (CH), 123.1 (CH), 120.0 (C0), 115.5 (C0), 109.4 (C0), 108.7
(C0), 96.6 (CH), 72.2 (CH), 71.0 (CH), 70.7 (CH), 66.9 (CH), 49.9 (2CH2), 49.6
(CH2), 33.7 (CH2), 31.9 (CH2), 30.0 (CH2), 29.4 (CH2), 27.4, 27.0, 26.3, 26.2,
25.1, 24.8, 24.5, 23.1, 22.8 (5CH2 and 4CH3); HRMS-ESI: calcd for [M-H]+
554.3588, found 554.3587.
4.2. AChE and BuChE inhibition assay
Swiss Webster mice provided by the Central Animal Laboratory of the
Oswaldo Cruz Foundation/RJ (Fiocruz) were used in this study. Animals were
held in light/dark cycle of 12/12 hours and fed ad-libitum. All experiments were
performed in accordance with the Ethics Committee for Animal Use of Fiocruz
(CEUA, license number L044/2015 with the additive license LA-010/2016). After
euthanasia, brain and blood were collected and used for the preparation of
biological fractions used in enzymatic assays.
A cerebral homogenate was prepared in distilled water then centrifuged at
20,000 G for 60 minutes at 4 ° C. The supernatant was discarded and the pellet
resuspended in 2 mL of 1% Triton X-100. After, this fraction was again
centrifuged at 14000 rpm (20800 G) for 90 minutes. The pellet was discarded
and the supernatants were used for enzymatic assays.
After allow the whole blood to clot at room temperature, the sample was
centrifuged at 2,000 x g for 10 minutes in a refrigerated centrifuge (40C). The
resulting supernatant, named serum fraction, was maintained in polypropylene
tubes until the enzymatic assays.
The activities of the AChE enzyme in the brain and BuChE in the serum
of mice were determined using a modified Ellman method.[68] Both brain
extract and blood serum were added to their respective reaction medium
27

containing sodium phosphate buffer 0.1 M pH 7.5 in a 96-well plate, as well as
the synthetic compounds to be analyzed. Control groups did not have the
addition of compounds. DTNB (5,5'-dithio-bis-(2-nitrobenzoic acid)) in the
concentration of 0.32 mM was added and the reaction medium preincubated at
25 °C for 10 minutes. The enzymatic assays were initiated by addition of the
substrates, acetylthiocholine iodide or butyrylthiocholine iodide at a final
concentration of 1.5 mM, and a further incubation was carried out for 10
minutes. After incubation, optical density was measured at wavelength 412 nm
and blank wells (without enzymatic activity) had their values discounted.
Compounds were first solubilized in 100% Dimethyl sulfoxide (DMSO).
Afterwards, they were diluted again in Milli-Q® water and tested against
inhibition on the enzymatic activities at the final concentrations ranging from
0.01 to 10,000 nM. Data analysis was performed using GraphPad Prism version
6.0 (GraphPad Software, Inc., San Diego, CA).
4.3. Molecular Modelling
We selected three representative conformations of AChE following an
ensemble docking strategy adopted on our previous study [39] to consider the
significant conformational changes observed on the peripheral anionic site
(PAS).[69] This methodology consists of docking compounds into each
representative conformation of the receptor, aiming to include implicitly the
protein flexibility.[70,71] The selected structures in this work were 1ZGC
(Torpedo californica),[63] 2CKM (Torpedo californica),[41] and 1Q84 (Mus
musculus).[72] Structures from Homo sapiens were not selected since they are
complexed with smaller inhibitors when compared with the compounds tested
herein. All inhibitors from the three representative conformations of AChE (i)
interact with the CAS, (ii) interact with the PAS, and (iii) are similar to the tacrine
inhibitor. No water molecules were considered in this work since they
demonstrated not to be important to predict experimental binding modes in our
28

redocking studies (results not shown). For the BuChE, we selected the
conformation complexed with the largest inhibitor 6QS (PDB ID 5K5E).[73]
Three-dimensional structures of the compounds were drawn using
ChemBio3D Ultra 14 Suite (PerkinElmer, Waltham, MA, USA, 2014) and
isomers, protonation states, and tautomer’s of the ligands were manually set
using Maestro (Maestro, Schrödinger, LLC, New York, NY, 2014). The AChE
and BuChE structures were taken from our previous study (Ceschi et al., 2016)
[39] which were prepared with Protein Preparation Wizard tool (Schrödinger
Suite 2014-1 Protein Preparation Wizard; Impact version 6.2, Schrödinger, LLC,
New York, NY, USA, 2014). The protonation states of the amino acid residues
were predicted using PROPKA with pH = 7 while respecting the protonation
state of the active site residues already described in the literature: Glu202 and
Glu327 negatively charged and neutral His440:ND1. Tacrine moiety of all
inhibitors was predicted to be protonated in the previous study according to the
Epik prediction.[74] This protonation state is essential since tacrine moiety
donates a hydrogen bond to the carbonyl oxygen of the catalytic His440 main
chain.[63] Optimization of the hydrogen bond network between protein and
inhibitor was performed to adjust the orientation of the hydrogen atoms,
followed by energy minimization with heavy atoms fixed.
In this work, the docking experiments were performed with the molecular
docking program GOLD [75] considering 200% of efficiency – indicated for
highly flexible ligands, setting early termination disabled and using ChemPLP to
score the ligand poses and to rank the different compounds. We also redocked
the co-crystallized ligands into their respective AChE and BuChE conformation
to validate the docking protocol adopted herein.
Acknowledgements
29